Skip to main content

Full text of "Raw_Thought-txt"

See other formats

Google and the Gradient
For a long time it seemed like everything I heard about Google was even cooler than the last. Wow, it’s a great search engine! Wow, they’re not sleazy like other companies! Wow, they treat their hackers well! Wow, theyarehackers!
The feeling peaked sometime last year when I was almost rolling on the floor hoping to work at Google. And I do mean peaked. Everything I’ve heard since then has been downhill, each time I hear about it Google seems less cool. I’m not saying the company is imminently doomed or that you should sell your shares, but I definitely don’t think it’s going to get any cooler.
Google is run like a socialist state. Its citizens are treated extremely well. There’s free food, free doctors, free massages, free games, a limited workweek, etc. There are ministries to give projects free promotion and support. The government tries to avoid getting too much in people’s lives. And Google is always coming up with more perks to give away. (There’s also a strong class hierarchy, with abused temps and powerful acquirees.)
The problem with a system like this is that it’s necessarily a bubble. Everybody inside gets treated grandly, but the outside world gets nothing. Indeed, because of Google’s notorious secrecy, they barely even get to talk to the people inside. A friend who’s a prominent free software developer says that every community member who’s joined Google has stopped contributing to public projects. It’s so bad, he says, that they’re thinking of banning Google from buying a booth at their next conference. They can’t afford to lose any more developers.
Which means that Google has to be careful about who they hire, but since they’re growing so fast they need tohire people as quickly as possible. It’s an impossible bind — you can hire lots of people or you can hire really good people, but even a company as prominent as Google is going to have a hard time doing both.
The solution, of course, is to pop the bubble. There’s no reason being part of Google has to be a binary decision. Google has a wide variety of resources and while there are some they can’t really give away to everyone (e.g. massages), there are others that should be easy (e.g. servers). Unfortunately for them, Google’s mindset is so obviously set that this will never happen. Even a company as woeful as Amazon is already kicking their but in this space, giving away storage space and computer power, with more in the works.
But let’s imagine you had the resources to do this right, what would you do? (I feel like I’m giving away a valuable secret here, but since nobody listens to me anyway, I doubt it will make any difference.) The right thing is to build not a bubble, with it’s binary in-or-out choice, but to build a gradient, with shades of resources you make available as people achieve success.
So you have this organization dedicated to building cool web apps. The first thing you do is you start giving away free food in the middle of San Francisco. You have a nice cozy area with tables and bathrooms and Wi-Fi and anyone interested in starting a web site is encouraged to drop by and hang out. There they can eat, chat, hack, get feedback, get suggestions, get help.
Then you give them free hosting. Servers and bandwidth are cheap, good projects are invaluable. But not only will you host their app for free, throwing in servers to scale it as necessary, but you’llpaythem for the privilege of hosting. Indeed, you’ll pay them proportionately to the amount of traffic they get, in exchange for the right to run ads on it someday.
So now you’ve got all the bright, smart young things who want to start companies starting them on your servers, with clear and unambiguous incentives: get traffic, get paid. They don’t need to worry about impressing anyone with their idea; anyone can use the hosting. And they don’t need to sell out to investors anymore; as their traffic grows, you’ll already be giving them the cash to grow the business.
Most of these sites, of course, will probably be failures. But who cares? Sites that don’t get much traffic don’t use up much in the way of resources. Meanwhile, a couple of the sites will actually take off. So what do you do with those?Give them more resources.
Put your promotional team behind them to spread the word about the ideas. Have your web designers, database jockeys, and JavaScript hotshots help them fix up the site. Encourage promising young programmers interested in helping out with something to write a feature or two.
And — this is where the gradient comes in — as they become more successful, you give them more resources. Let them move into the apartment building above to food/hangout space, so they can get more facetime with fellow successful hackers. Give them free offices to work in. Provide free massages and exercise equipment. Have your PR team set up interviews with the major media. Integrate their site with your other sites. Plus, of course, they’re getting paid more for more traffic the whole time.
Some of the sites will be huge hits, another YouTube or Facebook. The founders will be raking in millions from the traffic. And at some point, they’ll get tired of running the site and they’ll let it go. You’ll be there to take it over, slap some ads on it to recoup the investment, and give it to some new, junior developers to maintain and improve. And the cycle continues.
(Bonus for the truly adventurous: run the whole thing as a non-profit and have all the applications involved be open source.)
A bubble like Google can hire only so many people and there’s no way of picking only the ones which will be successes. But everyone can be part of a gradient and the successes simply rise to the top. I know which one I’d work for.
Thanks to Emmett Shear for discussion and suggestions.
You should follow me on twitterhere.
October 26, 2006